arXiv subjects

Yanbei Chen

Publications and source records attributed to Yanbei Chen.

At least 73 records · Page 4Linked to original sources

arXiv2023

Weak Cosmic Censorship and Second Law of Black Hole Thermodynamics in Higher Derivative Gravity

Infalling matter may destroy a black hole and expose the naked singularity. Thus, Penrose proposed the weak cosmic censorship conjecture to avoid such a possibility. On the other hand, if the black hole is not destroyed by infalling matter, from the second law of black hole thermodynamics, the black hole entropy should increase due to the information carried by the infalling matter. In this work, we demonstrate by examples of perturbative near-extremal black holes in higher derivative gravity theories that the second law implies weak cosmic censorship. We also compare our proposal to the one developed by Sorce and Wald based on the first law of black hole thermodynamics and show that the latter fails to yield weak cosmic censorship in such cases. Finally, we give proof of our proposal for generic gravity theories.

hep-th↗

arXiv2023

Exoplanets Around Red Giants: Distribution and Habitability

As the search for exoplanets continues, more are being discovered orbiting Red Giant stars. We use current data from the NASA Exoplanet Archive to investigate planet distribution around Red Giant stars and their presence in the host's habitable zone. As well, we update the power law relation between planet mass and stellar radius found in previous studies and provide more detailed investigations on this topic. Ten Red Giant-hosted exoplanets are found to be in the optimistically calculated habitable zone, five of which are in a more conservatively calculated habitable zone. We believe additional exoplanets can be found in habitable zones around Red Giants using the direct imaging and other methods, along with more powerful detection instrumentation.

astro-ph.EP↗

arXiv2023

Rogue echoes from exotic compact objects

Binary systems containing exotic compact objects may emit repeated bursts of gravitational waves (GWs) following coalescence. Such GW echoes would provide a clear signature of new physics, but searches for them have not yielded a convincing detection. Here we argue that the typical time delay between a GW event and its echoes is much greater than generally expected, due to long propagation times through objects that mimic black holes. We provide a simple recipe for computing the time delay and several examples. These time delays can be billions of years, resulting in rogue echoes that are not correlated with GW events and evade all current constraints. They would be detectable only by searches for individual echoes or GW bursts.

gr-qc↗

arXiv2023

Multiband Gravitational Wave Cosmography with Dark Sirens

Gravitational waves might help resolve the tension between early and late Universe measurements of the Hubble constant, and this possibility can be enhanced with a gravitational wave detector in the decihertz band as we will demonstrate in this study. Such a detector is particularly suitable for the multiband observation of stellar-mass black hole binaries between space and ground, which would significantly improve the source localization accuracy thanks to a long baseline for timing triangulation, hence promoting the "dark siren" cosmology. Proposed decihertz concepts include DECIGO/B-DECIGO, TianGO, and others. We consider here the prospects of multiband observation of dark siren binaries with a variety of network configurations. We find that a multiband observation can uniquely identify a black hole binary to a single galaxy to a cosmological distance, and thus a dark siren behaves as if it had an electromagnetic counterpart. Considering only fully localized dark sirens, we use a Fisher matrix approach to estimate the error in the Hubble constant and matter density parameter. We find that a decihertz detector substantially improves our ability to measure cosmological parameters because it enables host galaxies to be identified out to a larger distance without the systematics from statistical techniques based on comparing the population distribution.

gr-qc↗

arXiv2023

ScaleDet: A Scalable Multi-Dataset Object Detector

Multi-dataset training provides a viable solution for exploiting heterogeneous large-scale datasets without extra annotation cost. In this work, we propose a scalable multi-dataset detector (ScaleDet) that can scale up its generalization across datasets when increasing the number of training datasets. Unlike existing multi-dataset learners that mostly rely on manual relabelling efforts or sophisticated optimizations to unify labels across datasets, we introduce a simple yet scalable formulation to derive a unified semantic label space for multi-dataset training. ScaleDet is trained by visual-textual alignment to learn the label assignment with label semantic similarities across datasets. Once trained, ScaleDet can generalize well on any given upstream and downstream datasets with seen and unseen classes. We conduct extensive experiments using LVIS, COCO, Objects365, OpenImages as upstream datasets, and 13 datasets from Object Detection in the Wild (ODinW) as downstream datasets. Our results show that ScaleDet achieves compelling strong model performance with an mAP of 50.7 on LVIS, 58.8 on COCO, 46.8 on Objects365, 76.2 on OpenImages, and 71.8 on ODinW, surpassing state-of-the-art detectors with the same backbone.

cs.CV↗

arXiv2023

Perturbations of spinning black holes beyond General Relativity: Modified Teukolsky equation

The detection of gravitational waves from compact binary mergers by the LIGO/Virgo collaboration has, for the first time, allowed for tests of relativistic gravity in the strong, dynamical and nonlinear regime. Outside Einstein's relativity, spinning black holes may be different from their general relativistic counterparts, and their merger may then lead to a modified ringdown. We study the latter and, for the first time, derive a modified Teukolsky equation, i.e., a set of linear, decoupled differential equations that describe dynamical perturbations of non-Kerr black holes for the radiative Newman-Penrose scalars $Ψ_0$ and $Ψ_4$. We first focus on non-Ricci-flat, Petrov type D black hole backgrounds in modified gravity, and derive the modified Teukolsky equation through direct decoupling and through a new approach, proposed by Chandrasekhar, that uses certain gauge conditions. We then extend this analysis to non-Ricci-flat, Petrov type I black hole backgrounds in modified gravity, assuming they can be treated as a linear perturbation of Petrov type D, black hole backgrounds in GR by generalizing Chandrasekhar's approach, and derive the decoupled modified Teukolsky equation. Our work lays the foundation to study the gravitational waves emitted in the ringdown phase of black hole coalescence in modified gravity for black holes of any spin. Our work can also be extended to compute gravitational waves emitted by extreme mass-ratio binary inspirals in modified gravity.

gr-qc↗

arXiv2023

Quantum Gravity Background in Next-Generation Gravitational Wave Detectors

We study the effects of geontropic vacuum fluctuations in quantum gravity on next-generation terrestrial gravitational wave detectors. If the VZ effect proposed in Ref. [1], as modeled in Refs. [2, 3], appears in the upcoming GQuEST experiment, we show that it will be a large background for astrophysical gravitational wave searches in observatories like Cosmic Explorer and the Einstein Telescope.

gr-qc↗

arXiv2023

Using rational filters to uncover the first ringdown overtone in GW150914

There have been debates in the literature about the existence of the first overtone in the ringdown of GW150914. We develop a novel Bayesian framework to reanalyze the data of this event, by incorporating a new technique, the "rational filter" that can clean particular modes from the ringdown signal. We examine the existence of the first overtone in GW150914 from multiple novel perspectives. First, we confirm that the estimates of the remnant black hole mass and spin are more consistent with those obtained from the full IMR signal when including the first overtone at an early stage of the ringdown (right after the inferred signal peak); such improvement fades away at later times. Second, we formulate a new way to compare the ringdown models with and without the first overtone by calculating the Bayes factor at different times during the ringdown. We obtain a Bayes factor of 600 at the time when the signal amplitude reaches its peak. The Bayes factor decreases sharply when moving away from the peak time and eventually oscillates around a small value when the overtone signal is expected to have decayed. Third, we clean the fundamental mode from the ringdown of GW150914 and estimate the amplitudes of the modes using the filtered data with MCMC. The inferred amplitude of the fundamental mode is ~0 whereas the amplitude of the first overtone remains almost unchanged, implying that the filtered data is consistent with a first-overtone-only template. Similarly, if we remove the first overtone from the GW150914 data, the filtered data are consistent with a fundamental-mode-only template. Finally, after removing the fundamental mode, we use MCMC to infer the remnant black hole mass and spin from the first overtone alone. We find the posteriors are still informative and consistent with those inferred from the fundamental mode.

gr-qc↗

arXiv2023

Black hole spectroscopy by mode cleaning

We formulate a Bayesian framework to analyze ringdown gravitational waves from colliding binary black holes and test the no-hair theorem. The idea hinges on mode cleaning -- revealing subdominant oscillation modes by removing dominant ones using newly proposed ${\it rational~filters}$. By incorporating the filter into Bayesian inference, we construct a likelihood function that depends only on the mass and spin of the remnant black hole (no dependence on mode amplitudes and phases) and implement an efficient pipeline to constrain the remnant mass and spin without Markov chain Monte Carlo (MCMC). We test ringdown models by cleaning combinations of different modes and evaluating the consistency between the residual data and pure noise. The model evidence and Bayes factor are used to demonstrate the presence of a particular mode and to infer the mode starting time. In addition, we design a hybrid approach to estimate the remnant black hole properties exclusively from a single mode using MCMC after mode cleaning. We apply the framework to GW150914 and demonstrate more definitive evidence of the first overtone by cleaning the fundamental mode. This new framework provides a powerful tool for black hole spectroscopy in future gravitational-wave events.

gr-qc↗

arXiv2023

Nonlinearities in Black Hole Ringdowns

The gravitational wave strain emitted by a perturbed black hole (BH) ringing down is typically modeled analytically using first-order BH perturbation theory. In this Letter we show that second-order effects are necessary for modeling ringdowns from BH merger simulations. Focusing on the strain's $(\ell,m)=(4,4)$ angular harmonic, we show the presence of a quadratic effect across a range of binary BH mass ratios that agrees with theoretical expectations. We find that the quadratic $(4,4)$ mode's amplitude exhibits quadratic scaling with the fundamental $(2,2)$ mode -- its parent mode. The nonlinear mode's amplitude is comparable to or even larger than that of the linear $(4,4)$ mode. Therefore, correctly modeling the ringdown of higher harmonics -- improving mode mismatches by up to 2 orders of magnitude -- requires the inclusion of nonlinear effects.

gr-qc↗

arXiv2023

Designing optimal linear detectors -- a bottom-up approach

This paper develops a systematic approach to realising linear detectors with an optimised sensitivity, allowing for the detection of extremely weak signals. First, general constraints are derived on a specific class of input-output transfer functions of a linear detector. Then a physical realization of transfer functions in that class is found using the quantum network synthesis technique, which allows for the inference of the physical setup directly from the input-output transfer function. By exploring a minimal realization which has the minimum number of internal modes, it is shown that the optimal such detectors are internal squeezing schemes. Then, investigating non-minimal realizations, which is motivated by the parity-time symmetric systems, a quantum non-demolition measurement is systematically recovered.

quant-ph↗

arXiv2022

Quasinormal-mode filters: a new approach to analyze the gravitational-wave ringdown of binary black-hole mergers

We propose two frequency-domain filters to analyze ringdown signals of binary black hole mergers. The first rational filter is constructed based on a set of (arbitrary) quasi-normal modes (QNMs) of the remnant black holes, whereas the second full filter comes from the transmissivity of the remnant black holes. The two filters can remove corresponding QNMs from original time-domain ringdowns, while changing early inspiral signals in a trivial way - merely a time and phase shift. After filtering out dominant QNMs, we can visualize the existence of various subdominant effects. For example, by applying our filters to a GW150914-like numerical relativity (NR) waveform, we find second-order effects in the (l = 4, m = 4), (l = 5, m = 4) and (l = 5, m = 5) harmonics; the spherical-spheroidal mixing mode in the (l = 2,m = 2) harmonic; and a mixing mode in the (l = 2,m = 1) harmonic due to a gravitational recoil. In another NR simulation where two component spins are anti-aligned with the orbital angular momentum, we also find retrograde modes. The filters are sensitive to the remnant properties (i.e., mass and spin) and thus have a potential application to future data analyses and parameter estimations. We also investigate the stability of the full filter. Its connection to the instability of QNM spectra is discussed.

gr-qc↗

arXiv2022

Distilling Knowledge from Self-Supervised Teacher by Embedding Graph Alignment

Recent advances have indicated the strengths of self-supervised pre-training for improving representation learning on downstream tasks. Existing works often utilize self-supervised pre-trained models by fine-tuning on downstream tasks. However, fine-tuning does not generalize to the case when one needs to build a customized model architecture different from the self-supervised model. In this work, we formulate a new knowledge distillation framework to transfer the knowledge from self-supervised pre-trained models to any other student network by a novel approach named Embedding Graph Alignment. Specifically, inspired by the spirit of instance discrimination in self-supervised learning, we model the instance-instance relations by a graph formulation in the feature embedding space and distill the self-supervised teacher knowledge to a student network by aligning the teacher graph and the student graph. Our distillation scheme can be flexibly applied to transfer the self-supervised knowledge to enhance representation learning on various student networks. We demonstrate that our model outperforms multiple representative knowledge distillation methods on three benchmark datasets, including CIFAR100, STL10, and TinyImageNet. Code is here: https://github.com/yccm/EGA.

cs.CV↗

arXiv2022

Displacement-noise-free interferometeric gravitational-wave detector using unidirectional neutrons with four speeds

For further gravitational wave (GW) detections, it is significant to invent a technique to reduce all kinds of mirror displacement noise dominant at low frequencies for ground-based detectors. The neutron displacement-noise-free interferometer (DFI) is one of the tools to reduce all the mirror displacement noise at lower frequencies. In this paper, we describe a further simplified configuration of a neutron DFI in terms of neutron incidence direction. In the new configuration, neutrons enter the interferometer with unidirectional incidence at four speeds as opposed to two bidirectional incidences of opposite directions at two speeds as reported previously. This simplification of the neutron DFI is significant for proof-of-principle experiments.

gr-qc↗

arXiv2022

Cross-Modal Fusion Distillation for Fine-Grained Sketch-Based Image Retrieval

Representation learning for sketch-based image retrieval has mostly been tackled by learning embeddings that discard modality-specific information. As instances from different modalities can often provide complementary information describing the underlying concept, we propose a cross-attention framework for Vision Transformers (XModalViT) that fuses modality-specific information instead of discarding them. Our framework first maps paired datapoints from the individual photo and sketch modalities to fused representations that unify information from both modalities. We then decouple the input space of the aforementioned modality fusion network into independent encoders of the individual modalities via contrastive and relational cross-modal knowledge distillation. Such encoders can then be applied to downstream tasks like cross-modal retrieval. We demonstrate the expressive capacity of the learned representations by performing a wide range of experiments and achieving state-of-the-art results on three fine-grained sketch-based image retrieval benchmarks: Shoe-V2, Chair-V2 and Sketchy. Implementation is available at https://github.com/abhrac/xmodal-vit.

cs.CV↗

arXiv2022

Interferometer Response to Geontropic Fluctuations

We model vacuum fluctuations in quantum gravity with a scalar field, characterized by a high occupation number, coupled to the metric. The occupation number of the scalar is given by a thermal density matrix, whose form is motivated by fluctuations in the vacuum energy, which have been shown to be conformal near a light-sheet horizon. For the experimental measurement of interest in an interferometer, the size of the energy fluctuations is fixed by the area of a surface bounding the volume of spacetime being interrogated by an interferometer. We compute the interferometer response to these "geontropic" scalar-metric fluctuations, and apply our results to current and future interferometer measurements, such as LIGO and the proposed GQuEST experiment.

gr-qc↗

arXiv2022

Quantum scrambling in a toy model of photodetectors

Quantum measurement is a process that involves the interaction between a quantum system and a macroscopic measurement apparatus containing many degrees of freedom. The photodetector is such an apparatus with many electrons interacting with the incoming quantum photon. Therefore the incoming photon will spread and get scrambled in the photodetector, that is, the operator of the initial incoming local photons will grow and becomes highly non-local through the interaction process. Investigating this scrambling process in detail is useful for understanding the interaction between the quantum system and the measurement apparatus. In this paper, we study the quantum scrambling process in an effective toy model of photodetectors in three different physical scenarios, by numerically simulating the evolution of the out-of-time correlators (OTOC). In particular, the integrability of the effective model is explored through level spacing statistics, and the effect of the spatially/temporarily distributed disorders on the system evolution is carefully investigated using the OTOC. Looking into these detailed dynamical processes paves the way to the quantum simulation and manipulation of photondetector, which would provide insights into understanding the wave-function collapse processes.

quant-ph↗

arXiv2022

Semi-Supervised and Unsupervised Deep Visual Learning: A Survey

State-of-the-art deep learning models are often trained with a large amount of costly labeled training data. However, requiring exhaustive manual annotations may degrade the model's generalizability in the limited-label regime. Semi-supervised learning and unsupervised learning offer promising paradigms to learn from an abundance of unlabeled visual data. Recent progress in these paradigms has indicated the strong benefits of leveraging unlabeled data to improve model generalization and provide better model initialization. In this survey, we review the recent advanced deep learning algorithms on semi-supervised learning (SSL) and unsupervised learning (UL) for visual recognition from a unified perspective. To offer a holistic understanding of the state-of-the-art in these areas, we propose a unified taxonomy. We categorize existing representative SSL and UL with comprehensive and insightful analysis to highlight their design rationales in different learning scenarios and applications in different computer vision tasks. Lastly, we discuss the emerging trends and open challenges in SSL and UL to shed light on future critical research directions.

cs.CV↗